[ckan-dev] [ckan-discuss] Package Resources Proposal

Jonathan Gray jonathan.gray at okfn.org
Wed Feb 2 15:47:56 UTC 2011


This looks great!

One small question:

On Wed, Feb 2, 2011 at 3:57 PM, James Gardner <james at 3aims.com> wrote:
>  openness
>      A rating of how open the resource is (5 stars of openness), 0
>      if the resource currently isn't present.

I presume that this will not supersede 'license' field? I can
understand how this is probably needed/wanted for data.gov.uk, but I
wonder how this is being operationalised? Is it straightfoward? Can it
be calculated on the basis of other fields or does it have to be
entered manually? If manually entered, who is doing this for
data.gov.uk datasets, and are there any issues in applying the stars?

For what its worth, I think the existing 'bulk download + open license
= open' formula (= can I grab the stuff, can I reuse the stuff) is
pretty good, and we (will) have all kinds of nice ways of browsing for
datasets in different formats, etc, if users have particular
requirements in this request.

As far as I can tell the weighting of the 5 stars rating is:

  * agnostic about legal openness (which is very important for
prospective reusers to know about)
  * strongly biased towards open formats (whereas arguably machine
readability is the most important thing, and e.g. if the material is
openly licensed in XLS, then it doesn't take much for someone to
create and contribute back a CVS/XML/RDF version)
  * strongly biased towards Linked Data technologies, which the OKF is
not committed to *per se*, in the way that it is committed to machine
readable data that is open as in opendefinition.org (I *think* as an
organisation we are a bit more format/technology agnostic than the 5
stars are)

Its great to see this being implemented as a feature for those who
want it, but I'd be a bit cautious about rolling it out across all
CKAN instances and would want to think carefully how it interacts with
our more general work on developing standards in relation to open
knowledge and open data (such as opendefinition.org). Though in
practise I'm sure this will not be a concern (at least in the UK where
the Open Government License is widely, but not *quite* ubiquitously,
used), in principle one could imagine a 5 star dataset under a custom
non-commercial license, with all kinds of interoperability problems.
Which is not what open government data should be about! ;-)

-- 
Jonathan Gray

Community Coordinator
The Open Knowledge Foundation
http://blog.okfn.org

http://twitter.com/jwyg
http://identi.ca/jwyg




More information about the ckan-dev mailing list